[22p-52A-1~17] 23.1 Joint Session N "Informatics"

3:00 PM - 3:15 PM JST(6:00 AM - 6:15 AM UTC)

[22p-52A-8] Graph neural networks learning for superconducting temperatures on partially-substituted materials

〇Kensei Terashima¹, Taku Tou^1,2, Yoshihiko Takano^1,3 (1.NIMS, 2.Tokyo Univ. of Sci., 3.Univ. of Tsukuba)

Keywords:

machine learning

We conducted Graph Neural Network (GNN) learning based on ALIGNN for the transition temperature (Tc) of superconductors, including partially substituted samples. This approach demonstrates comparable or superior accuracy to cases without partial substitution. Moreover, it suggests an expanded search space when using known-structural materials.